Integrated Drug and Medical Care: Cost and Effectiveness This is a competing continuation of a study comparing outcomes of individuals randomized to Integrated versus non-integrated (Usual Care) medical and substance abuse treatment. The study addressed the integration of drug abuse treatment and primary medical care, an important delivery system issue. Although individuals entering substance abuse treatment have a high prevalence of medical disorders, traditionally the two service systems have been independent and uncoordinated. The proposed study uses a conceptual model drawn from our earlier work and from longitudinal literature to identify the trajectory determinants of substance use, medical, and mental health outcomes over 7 years. This is relevant to theoretical and clinical research in epidemiology and health services, where substance abuse problems are increasingly viewed as chronic and relapsing. The study follows an intent-to-treat sample of individuals first interviewed during admission to substance abuse treatment in a large group-model health maintenance organization (HMO). The original study found that individuals with medical conditions related to substance abuse who were randomized to integrated care had higher rates of abstinence than those randomized to usual care. Integration of carewas also cost-effective, and produced a cost offset, when comparing the 12 months before and after treatment, with a matched cohort from the health plan. We propose follow-up interviews with the original sample at 3 year intervals (4 and 7 years post intake), and use of the HMO's automated databases to examine continuous measures of utilization of medical services and cost over the 7 year follow-up period. The conceptual model examines the roles played by individual characteristics (including medical conditions and psychiatric status); index treatment factors (integrated versus usual care, length of stay) and readmissions; and extra-treatment factors (e.g., routine primary care and prevention services, participation in self help groups, and changes in social networks) in altering the 7-year course of substance use. Latent curve analysis techniques will be used to examine the shape of trajectories of substance use and medical outcomes, and their determinants, and parallel analyses to examine medical utilization and cost outcomes over time. The multiple data points are important in examining substance use patterns, medical status, and utilization over time, and their 7-year coverage provides sufficient time to examine long-term effects of pretreatment medical conditions, index treatment episodes and cumulative effects of treatment, other medical services, and lifestyle changes. An additional strength is the organizationalsetting of a "real world" HMO, its comprehensive automated databases, andthe potential for dissemination to other regions.